Commit bb6fbc45 authored by Trond Myklebust's avatar Trond Myklebust

NFS: Avoid a deadlock in nfs_release_page

J.R. Okajima reports the following deadlock:

INFO: task kswapd0:305 blocked for more than 120 seconds.
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
kswapd0       D 0000000000000001     0   305      2 0x00000000
 ffff88001f21d4f0 0000000000000046 ffff88001fdea680 ffff88001f21c000
 ffff88001f21dfd8 ffff88001f21c000 ffff88001f21dfd8 ffff88001f21dfd8
 ffff88001fdea040 0000000000014c00 0000000000000001 ffff88001fdea040
Call Trace:
 [<ffffffff8146155d>] io_schedule+0x4d/0x70
 [<ffffffff810d2be5>] sync_page+0x65/0xa0
 [<ffffffff81461b12>] __wait_on_bit_lock+0x52/0xb0
 [<ffffffff810d2b80>] ? sync_page+0x0/0xa0
 [<ffffffff810d2b64>] __lock_page+0x64/0x70
 [<ffffffff81070ce0>] ? wake_bit_function+0x0/0x40
 [<ffffffff810df1d4>] truncate_inode_pages_range+0x344/0x4a0
 [<ffffffff810df340>] truncate_inode_pages+0x10/0x20
 [<ffffffff8112cbfe>] generic_delete_inode+0x15e/0x190
 [<ffffffff8112cc8d>] generic_drop_inode+0x5d/0x80
 [<ffffffff8112bb88>] iput+0x78/0x80
 [<ffffffff811bc908>] nfs_dentry_iput+0x38/0x50
 [<ffffffff811285f4>] dentry_iput+0x84/0x110
 [<ffffffff811286ae>] d_kill+0x2e/0x60
 [<ffffffff8112912a>] dput+0x7a/0x170
 [<ffffffff8111e925>] path_put+0x15/0x40
 [<ffffffff811c3a44>] __put_nfs_open_context+0xa4/0xb0
 [<ffffffff811cb5d0>] ? nfs_free_request+0x0/0x50
 [<ffffffff811c3b0b>] put_nfs_open_context+0xb/0x10
 [<ffffffff811cb5f9>] nfs_free_request+0x29/0x50
 [<ffffffff81234b7e>] kref_put+0x8e/0xe0
 [<ffffffff811cb594>] nfs_release_request+0x14/0x20
 [<ffffffff811cf769>] nfs_find_and_lock_request+0x89/0xa0
 [<ffffffff811d1180>] nfs_wb_page+0x80/0x110
 [<ffffffff811c0770>] nfs_release_page+0x70/0x90
 [<ffffffff810d18ee>] try_to_release_page+0x5e/0x80
 [<ffffffff810e1178>] shrink_page_list+0x638/0x860
 [<ffffffff810e19de>] shrink_zone+0x63e/0xc40

We can fix this by making the call to put_nfs_open_context() happen when we
actually remove the write request from the inode (which is done by the
nfsiod thread in this case).
Signed-off-by: default avatarTrond Myklebust <Trond.Myklebust@netapp.com>
Cc: stable@kernel.org
parent b4d2314b
...@@ -112,12 +112,10 @@ void nfs_unlock_request(struct nfs_page *req) ...@@ -112,12 +112,10 @@ void nfs_unlock_request(struct nfs_page *req)
*/ */
int nfs_set_page_tag_locked(struct nfs_page *req) int nfs_set_page_tag_locked(struct nfs_page *req)
{ {
struct nfs_inode *nfsi = NFS_I(req->wb_context->path.dentry->d_inode);
if (!nfs_lock_request_dontget(req)) if (!nfs_lock_request_dontget(req))
return 0; return 0;
if (req->wb_page != NULL) if (req->wb_page != NULL)
radix_tree_tag_set(&nfsi->nfs_page_tree, req->wb_index, NFS_PAGE_TAG_LOCKED); radix_tree_tag_set(&NFS_I(req->wb_context->path.dentry->d_inode)->nfs_page_tree, req->wb_index, NFS_PAGE_TAG_LOCKED);
return 1; return 1;
} }
...@@ -126,10 +124,10 @@ int nfs_set_page_tag_locked(struct nfs_page *req) ...@@ -126,10 +124,10 @@ int nfs_set_page_tag_locked(struct nfs_page *req)
*/ */
void nfs_clear_page_tag_locked(struct nfs_page *req) void nfs_clear_page_tag_locked(struct nfs_page *req)
{ {
if (req->wb_page != NULL) {
struct inode *inode = req->wb_context->path.dentry->d_inode; struct inode *inode = req->wb_context->path.dentry->d_inode;
struct nfs_inode *nfsi = NFS_I(inode); struct nfs_inode *nfsi = NFS_I(inode);
if (req->wb_page != NULL) {
spin_lock(&inode->i_lock); spin_lock(&inode->i_lock);
radix_tree_tag_clear(&nfsi->nfs_page_tree, req->wb_index, NFS_PAGE_TAG_LOCKED); radix_tree_tag_clear(&nfsi->nfs_page_tree, req->wb_index, NFS_PAGE_TAG_LOCKED);
nfs_unlock_request(req); nfs_unlock_request(req);
...@@ -142,16 +140,22 @@ void nfs_clear_page_tag_locked(struct nfs_page *req) ...@@ -142,16 +140,22 @@ void nfs_clear_page_tag_locked(struct nfs_page *req)
* nfs_clear_request - Free up all resources allocated to the request * nfs_clear_request - Free up all resources allocated to the request
* @req: * @req:
* *
* Release page resources associated with a write request after it * Release page and open context resources associated with a read/write
* has completed. * request after it has completed.
*/ */
void nfs_clear_request(struct nfs_page *req) void nfs_clear_request(struct nfs_page *req)
{ {
struct page *page = req->wb_page; struct page *page = req->wb_page;
struct nfs_open_context *ctx = req->wb_context;
if (page != NULL) { if (page != NULL) {
page_cache_release(page); page_cache_release(page);
req->wb_page = NULL; req->wb_page = NULL;
} }
if (ctx != NULL) {
put_nfs_open_context(ctx);
req->wb_context = NULL;
}
} }
...@@ -165,9 +169,8 @@ static void nfs_free_request(struct kref *kref) ...@@ -165,9 +169,8 @@ static void nfs_free_request(struct kref *kref)
{ {
struct nfs_page *req = container_of(kref, struct nfs_page, wb_kref); struct nfs_page *req = container_of(kref, struct nfs_page, wb_kref);
/* Release struct file or cached credential */ /* Release struct file and open context */
nfs_clear_request(req); nfs_clear_request(req);
put_nfs_open_context(req->wb_context);
nfs_page_free(req); nfs_page_free(req);
} }
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment